Stx2f producing-E. coli represent a serious problem in public health

Shiga toxins (Stxs) are the main virulence determinants associated with STEC. The genes encoding the Stxs are harboured by temperate lambdoid bacteriophages (O'Brien et al., 1984). Differences in

Figure 7.1 Representation of E. coli pangenome constituted of about 18,000 genes, of which 11% form the conserved part, the “core genome” (dark blue), 62% represents the “persistent genes” (blue), 26% can be considered “accessory genes” (pale blue).

Evolutionary forces induce acquisition and loss of genetic determinants, leading to an average of E. coli genome of about 4,700 genes (van Elsas et al., 2011).

Chapter 7

78

the DNA sequence of the stx genes led to classify the Stxs in two main antigenically diverse types, Stx1 and Stx2, which in turn are subdivided into three (Stx1a, Stx1c and Stx1d) and seven subtypes (Stx2a-g) respectively (Persson et al., 2007).

The most severe diseases in human have been frequently associated to the toxin subtype produced by the infecting strains (Friedrich et al., 2002). For long time STEC producing Stx2f subtype have been considered of scarce epidemiological relevance, as they were considered adapted to pigeons, their natural reservoir, and rarely associated with human infections. Therefore, cases of infections by Stx2f-producing E. coli have been long ignored and not efficiently surveyed.

Originally, only a few sporadic cases of mild diarrhoea associated to the Stx2f variant were reported in literature, but nowadays an increasing number of stx2f positive strains, isolated from human disease and belonging to a great variety of different serotypes, have been described (Prager et al., 2009, Friesema et al., 2014), suggesting that a more thorough assessment of the risk associated with this STEC population was needed.

We have carried out a whole genome characterization of E. coli strains producing the Stx2f subtype including isolates from diarrhoea, HUS and the natural reservoir, in order to provide evidence on the possible origin of the Stx2f-producing STEC causing HUS. Virulotyping of the isolates highlighted the existence of distinct subpopulations of these strains, which suggested a general variability of such STEC population was not identified previously. In particular, the isolates grouped based on their virulence genes content and with different degrees of overlapping with those present in the animal reservoir. The same analysis showed that the isolates from HUS cases displayed an asset of virulence genes additional to the stx resembling that of the more typical STEC producing Stx2 subtypes, such as the Stx2a, normally associated with HUS and belonging to serogroups such as O157, O26, O111, O103 and O145. Beside the chromosomally encoded genes governing the A/E lesion, and the related accessory factors such as the non-LEE encoded effectors (nle), encoded by the nleABC genes (Garmendia et al., 2005), the Stx2f-producing STEC from HUS also carried the OI-122 pathogenicity island (Karmali et al., 2003), and the genes ehxA, espP and katP genes, present on the large virulence plasmid first described in STEC O157 (pO157) (Beutin et al., 1990, Brunder et al., 1996, Brunder et al., 1997). Our results allowed us to speculate on the potential association of the Stx subtypes with the severity of the symptoms. As a matter of fact, our study indicated that even though the Stx2f was previously considered as being adapted to the pigeon host and not involved in the induction of severe disease in humans, it still has the potential to cause HUS when produced by an isolate able to efficiently colonize the human intestinal tract.

In this respect, it is important to note that the Stx2f-STEC from pigeons and mild diarrhoea specimens were all positive for the presence of the LEE locus, but the latter lacked of all the

Chapter 7

79

accessory determinants such as the nle and the genes conveyed by the OI-122 and pO157. At the same time the pigeon isolates showed an intermediate virulence genes asset with the presence of nle but did not possess the other two mentioned MGEs.

The compartmentalisation of the different subpopulations of Stx2f-producing STEC was also observed through the principal component analysis of virulence genes profile and a phylogenetic analysis conducted considering the whole genome SNPs showed that HUS Stx2f-producing strains clustered closer to other non-O157 STEC isolated from humans with severe disease, while the isolates from diarrhoea were more distantly related.

Whether the Stx2f-producing STEC isolated from HUS originated directly in the animal reservoir or through a mechanism of phage acquisition occurred in the environment and involving E. coli strains with a complete machinery for the induction of the A/E lesion, remains to be ascertained.

Further studies are also required to investigate the mechanism used by phages conveying stx2f genes to be released from the bacterial cell. In fact, these phages seem to lack most of the genes that regulate the switch between the lysogenic state and the lytic cycle typical of the lambdoid bacteriophages, as cro, cI, cII and cIII and we were not able to induce them in different experimental conditions (data not shown). Moreover, it will be interesting to assess the presence of free stx2f-paghes in the environment, in order to investigate whether these can be available to lysogenize other E. coli strain of non-pigeon sources.

7.2 “Bio-waste” as a new possible transmission pathway of pathogenic E. coli including STEC STEC infections occurs, mainly, via foodborne and waterborne routes, through inter-human cycles or direct contacts with STEC positive animals, or their environment. Contamination of water and soil used for agricultural use by STEC and other pathogenic E. coli is gaining increasing attention in public health.

Nowadays, the need to high crop with a reduced usage of water and expensive fertilizers to amend soils led to recycle bio-waste as a resource for energy, water and nutrients (Figure 7.2) (Nielsen, 2017). Such waste, defined also as Biosolid (BSO), refers to sewage sludge derived from biological or chemical treatment of industrial, municipal and zootechnical wastewater, manure from livestock, and compost from green wastes (Saveyn et al., 2014). Sewage sludge is subjected to both aerobic and anaerobic fermentation processes reaching high temperature, which, however, may not enough to remove potential hazards for human and animal health (Figure 7.2)

.

Chapter 7

80

We attempted to identify free stx2f-phages (See 7.1) in BSO samples used for the analysis described in Chapters 4 and 5 (data unpublished). To this purpose we selected BSO samples positive for the presence of stx2f gene subtype in Real Time PCR to infect a bacterial culture with the filtered supernatant. The results showed the presence of many lytic plaques (Figure 7.3), indicating the presence of infective phages. We tried to identify possible stx2f-phages among the plaques through plaque hybridisation with a probe specific for the stx2f gene, but with negative results.

Figure 7.2 Schematic representation of recovery cycle and reuse of wastewater http://www.billundbiorefinery.dk/en/.

Chapter 7

81

We have also explored the possibility that BSO may be a source of pathogenic E. coli including STEC. We detected the presence of virulence genes associated with both viral and bacterial enteric pathogens in various BSO samples intended for soil fertilization and grazing activities. The analysis, performed using Real Time PCR, gave positive results for virulence genes of human adenovirus, human norovirus and different diarrhoeagenic E. coli (DEC). The screening for pathogenic E. coli-associated genes was performed on the DNA extracted both directly from the specimens and from enriched cultures of the samples. As described in Chapter 4, some Real Time PCR signals, which were not identified in the DNA extracted directly from the samples, appeared after the enrichment for some of the virulence genes investigated, suggesting the presence of viable microbial cells. Five specimens negative for stx genes in direct screening, became positive after the enrichment, indicating that biosolids may indeed be a potential source of STEC and their use in agriculture may cause the transmission of such important human pathogens to soils, with a consequent risk of contamination of the food chain. As a matter of fact, it has been demonstrated

Figure 7.3 Agar plates containing phages plaques originated by lysis of the E. coli propagator strain upon infection with filtered supernatant derived from BSO samples.

A

B

C

Chapter 7

82

that viable E. coli O157:H7 strains can persist up to several months in the soil (Gagliardi et al., 2002).

Overall, the BSO samples analysed proved positive for the presence of virulence genes associated to all the E. coli pathotypes forming the DEC group. As a consequence, the genes investigated, all borne on mobile genetic elements (MGE), may be transferred to soils also in the absence of live bacterial cells and may be acquired by soil microorganisms, or by commensal species, including E.

coli, living in the intestine of grazing animals, contributing to the possible emergence of new pathogenic variants.

One of the main concerns regards the emergence of novel STEC strains with shuffled virulence features. Bacteriophages harbouring stx genes can be acquired by E. coli strains belonging to different pathogroups, favouring the possible emergence of pathogenic clones highly virulent to human (Frank et al., 2011, Tozzoli et al., 2014). As a matter of fact, the samples analysed were from different origin including wastewater plants treating both wastes from urban settlement and from livestock and, although we could not isolate the strain responsible for the PCR signals, probably due to the massive amount of microorganisms present in the specimens, the possibility exists that such practice of using BSO from wastewater treatment plants to fertilize agricultural soils may represent a risk for human health and should be better investigated.

In order to come to a deeper analysis of the BSO samples we used metagenomics as described in Chapter 5. In particular, we used this approach to define a complete microbiological profile of BSO specimens including samples from sludges and compost. Interestingly, the taxonomic profiling of the BSO microbiota has proven to be a good indicator of the origin of the different BSO, with samples from sludges grouping together and apart from those from compost.

The metagenomic analysis provided evidences that biosolids may act as a new environmental vehicle of important biological threats for human health. The presence of many genomics traits related with different diarrheagenic E. coli pathogroups was detected, including STEC, confirming previous results obtained for the same samples by Real Time PCR (Chapter 4). The holistic analysis conducted in this study through the shotgun metagenomics also allowed us to identify the presence in the BSO samples of gene sequences associated with the resistance to the main classes of antibiotics, as Fluoroquinolones, Beta-lactams, Streptothricin, Fosfomycin, Vancomycin and Methicillin, as well as with the resistance to the heavy metals. This finding is in line with the derivation of the BSO from sludges from wastewater treatment plants. As a matter of fact, antimicrobials are used both in the treatment of human diseases and in animal farming and thus our results may reflect the pressure applied to bacteria from human and animal origin.

Chapter 7

83

The metagenomic functional profiling of BSO may also provide an interesting tool for both retrospective and prospective analysis of emerging hazards. This may allow sound time-trends analysis to assess potential risks for animal and human health, as in the case of a screening for carbapenemase genes. Such genes were described to confer resistance to Carbapenems in environmental Enterobacteriaceae (Nordmann et al., 2011, Tzouvelekis et al., 2012). As a matter of fact, Carbapenems represent the last generation of Beta-lactams to fight multi-resistant bacteria responsible for serious infections at hospitals. However, there is rising evidence that wastewater effluents from hospital may play a role in disseminating Carbapenems resistance in the environment, through the genes-exchange between bacterial hosts in sewage (Ludden et al., 2017).

Although we did not carry out a specific analysis for this particular trait in the samples, it is important to note that one of the strength of the approach adopted is that the metagenomes are always available for further investigations and, given the huge amount of information contained into them, can be re-analysed with different bioinformatics procedures for specific aims.